Quantum Machine Learning in Log-based Anomaly Detection: Challenges and Opportunities

TL;DR

This work investigates the application of quantum machine learning to log-based anomaly detection within AIOps by introducing the QuLog framework that unifies data, quantum-classical hybrids, and comprehensive metrics. It quantumizes three classical LogAD models (DeepLog, LogAnomaly, LogRobust) and evaluates their performance across BGL, Spirit, and Thunderbird datasets, revealing that QML can match or exceed classical performance in some cases (notably Spirit) but is not universally superior and is highly sensitive to encoding, PQC design, and dataset characteristics. The results highlight significant parameter savings and efficient training for QML, while also underscoring limitations such as circuit complexity and robustness constraints on diverse data. The study charts practical directions for future QML-enabled LogAD, including optimized PQCs and quantum encodings, with an emphasis on real-world NISQ feasibility and scalable deployment.

Abstract

Log-based anomaly detection (LogAD) is the main component of Artificial Intelligence for IT Operations (AIOps), which can detect anomalous that occur during the system on-the-fly. Existing methods commonly extract log sequence features using classical machine learning techniques to identify whether a new sequence is an anomaly or not. However, these classical approaches often require trade-offs between efficiency and accuracy. The advent of quantum machine learning (QML) offers a promising alternative. By transforming parts of classical machine learning computations into parameterized quantum circuits (PQCs), QML can significantly reduce the number of trainable parameters while maintaining accuracy comparable to classical counterparts. In this work, we introduce a unified framework, \ourframework{}, for evaluating QML models in the context of LogAD. This framework incorporates diverse log data, integrated QML models, and comprehensive evaluation metrics. State-of-the-art methods such as DeepLog, LogAnomaly, and LogRobust, along with their quantum-transformed counterparts, are included in our framework.Beyond standard metrics like F1 score, precision, and recall, our evaluation extends to factors critical to QML performance, such as specificity, the number of circuits, circuit design, and quantum state encoding. Using \ourframework{}, we conduct extensive experiments to assess the performance of these models and their quantum counterparts, uncovering valuable insights and paving the way for future research in QML model selection and design for LogAD.

Figures (8)

• Figure 1: High-level depiction of hybrid quantum machine learning system for log analysis.
• Figure 2: Overview of the QuLog framework. The framework is structured incrementally, progressing from datasets to models and then to metrics.
• Figure 3: Four types of combine the disciplines of quantum computing and machine learning.
• Figure 4: Commonly used quantum gates and their corresponding unitary matrices.
• Figure 5: The training process of the classical-quantum hybrid machine learning model for log-based anomaly detection. The purple module represents the classical computer operation and the green part represents the quantum computer operation.